From Paths to Routes: A Method for Path Classification.

Area of Science:

• Animal behavior
• Computational biology
• Movement ecology

Background:

• Animals optimize routes for efficient navigation.
• Current methods for comparing animal paths (e.g., speed, distance, meandering) lack precision.
• Experimenter-based visual classification of paths introduces observer bias.

Purpose of the Study:

• To develop a novel, objective method for classifying animal movement paths based on trajectory similarity.
• To address limitations of existing path comparison techniques in understanding animal navigation.
• To provide advanced tools for trajectory analysis in diverse experimental contexts.

Main Methods:

• Utilized Dynamic Time Warping (DTW) and Fréchet Distance as key distance functions for trajectory comparison.
• Applied Monte Carlo Reference-Based Consensus Clustering to group similar trajectories based on calculated distances.
• Tested the method on bumblebee flight trajectories in a cluttered environment.

Main Results:

• The proposed method effectively classifies path similarities using DTW and Fréchet Distance.
• Consensus clustering successfully grouped trajectories based on the computed distance metrics.
• Demonstrated a robust approach for quantitative trajectory analysis.

Conclusions:

• The novel method provides a more accurate and less biased approach to analyzing animal path similarities.
• This technique enhances the understanding of route following and navigation strategies in animals.
• Offers versatile applications for trajectory analysis across various research paradigms.